The completed design can be found at : https://public.tableau.com/app/profile/qiyun.he/viz/DataVisMakeOver1/LFPR
1. Data Visualization Critique
The “Resident Labour Force Participation Rate by Age and Sex” file provides labor force participation rate(LFPR) at different age groups from year 1991-2021.
The intended visualization aims to show LFPR by age group for each year from year 2010-2021. The original design as shown below has clarify and aesthetic issues which will be discussed in the next section.
1.1. Critiques: Clarity
(a) Misleading labeling of bottom x-axis. and formatting of both x-axis and y-axis:
The label on the bottom x axis is misleading as it only shows 2015. As the graph needs to show LFPR from year 2010-2021 for each age group, this information should be easily identified by the user. As an improvement, a bar graph can be used to represent the LFPR value for each year under each age group. In this case, one color corresponds to a year, and user can identify the LFPR value for each year by referring to the legend.
(b) Improper labeling of upper x-axis and y-axis:
The first two labels “75 &O..” and “70& Ov..” for Age-group are not shown entirely,it would be better to flip the direction of label so that all labels for age group can be seen properly.
On the y-axis, the ‘Lfpr’ should be all capital letters. And it is not clear to the user what unit the axis is in. It would be better if the unit ‘%’ can be shown.
(b) Improper visualization header:
The title does not tell user clearly for what category the LFPR is shown for. It would be better if the title can mention that the LFPR is for each year from 2010-2021 of each age group.
Also, as ‘LFPR’ label of the y-axis is a short form for ‘Labour force participation rate’, it would be better to include it in a bracket beside the title so that user knows its meaning.
There is no subtitle, having a subtitle explaining the key observations from the graph can make it easier for user to understand the visualization.
(d) Redundant sorting of age groups according to LFPR
The age groups are sorted according to LFPR in an ascending order. This is redundant as the user may need to identify LFPR trends for people in working age or for elderly.
For example, in Singapore the typical ages for economic activity can be in the interval of 20-65, and the ages for elderly persons are from 65 years and over as referenced from page 2 of the website with link.
Therefore, displaying the age group in its original order would be better for user to identify the population of interest for analysis. To compare across age groups, statistics such as average LFPR, percentage change in LFPR for each year and overall percentage change from 2010-2021 can be added in.
1.2. Critiques: Aesthetics
(a) Redundant grid lines:
The shallow grid lines within each age group are redundant as it does not help in the visualization of any categorization of data and may be confusing to user when looking at it. Hence, it can be removed.
(b) Redundant color for each age group:
As each categorization of age group can be identified from the ‘Age-group’ axis label, is it redundant to assign different colors to different age groups and stated it in the legend. Also, assigning color to each age group does not help user to identify yearly change of LFPR. Hence, color should be used to differentiate years within each age group, and the legend can be changed to indicate the color corresponding to each color.
(c) Title does not stand out:
Title is not in bold, it will be better to format title as bold.
2. Sketch of Proposed Design
Below shows a sketch of the proposed design.
The proposed design will consist of bar graph for each year from 2010-2021 under each age group. The legend displays the color corresponding to each year. In this way, user can inspect the years as a whole, or only individual years according to color.
An additional section displaying the percentage change of LFPR from previous year is added, this helps user to understand the degree of fluctuation in LFPR for each age group.
To show the overall change from 2010-2021, another graph is created to display the % change, with red color indicating positive % change and blue color indicating negative % change, to give user an idea of the overall trend.
3. Data Visualization
3.1. Data Preparation
Step 1:
The data is downloaded from Ministry of Manpower website link titled ‘Resident Labour Force Participation Rate by Age and Sex’.
Under sheet ‘mrsd_Res_LFPR_2’, the data for each age group from year 1991 to 2021 is available.
-The visualization is only going to focus on total.
-As the visualization only focuses on LFPR per age group, the row ‘Total’ can be deleted.
-Also,the row ‘70 & over’ can be deleted as it can be separated into ‘70 to 74’ and ‘75 & Over’ and just retaining the later two rows are enough.
-The columns containing data for year 1991-2009 are deleted as the visualization only focuses on year 2010-2021.
After applying the above changes, create a separate sheet named ‘Total’ to store the data as shown below:
Import data into Tableau
Step 1:
Import sheet ‘Total’ into Tableau
Step 2:
Click on ‘Use Data Interpreter’ to clean the data. The columns are now with proper header ‘Age (Years)/Sex’ and years from 2010-2021.
Step 3:
Change column name ‘Age (Years)/Sex’ to ‘Age group’ by clicking on the ‘rename’ option and making the corresponding change.
Step 4:
Select all columns from 2010-2021, right click and select ‘pivot’:
In the resulting table, change the column name ‘Pivot Field Names’ to ‘Years’ and ‘Pivot Field Values’ to ‘LFPR(%)’.
3.2 Create graphs
3.2.1. Create the first graph
1. Create the basic parts of the graph
-Drag ‘Age Group’ and ‘Years’ to Columns, ‘LFPR%’ to
Rows, and drag ‘Years’ to Color, the following graph is displayed.
2. Adjust the labels and titles.
-Right click on the ‘Age Group / Years’ axis and select ‘Rotate Label’ to change the label orientation.
-Right click on the bottom x-axis and uncheck ‘Show Header’.
-Double click on ‘Sheet 1’ and type the title on the pop up box as shown below. Format the title as bold with font size 12.
-In the same pop up box, type subtitle under title, format the subtitle with font size 10. Align both title and subtitle to center.
-Right click on ‘Age Group’ and select ‘Format’. Click on ‘Font’ under ‘Default’ as shown below and set the size as 9. Apply the same process to ‘LFPR(%)’ label of the y-axis.
3. Add reference line
-Under Analytics, drag ‘Average Line’ to the graph and drop it under ‘Pane’ for first ‘SUM(LFPR(%))’.
The reference lines are now displayed on the graph:
-To show the values of reference lines, right click on any reference line, in the pop up box, select ‘Value’ for ‘Label’.
The values are now displayed:
3.2.2. Create the second graph
1. Create the second graph
-Drag ‘LFPR%’ to Rows again, right click on the second ‘SUM(LFPR(%))’ as shown below and select ‘Add Table Calculation’.
!
-In the pop up window, for ‘Calculation Type’, select
‘Percent Difference From’, and for ‘Compute Using’, select ‘Specific Dimensions’ and uncheck ‘Age Group’.
Above changes will produce a second graph as shown below:
-Click on ‘>13 nulls’ shown at the bottom right corner and select ‘Show data at default position’.
-Rename the sheet as ‘graph1-2’ by right clicking on ‘Sheet 1’ tab at the bottom and select ‘rename’.
2. Add in annotation and remove row grid lines
-Right click on the second graph and select ‘Annotate’->‘Area’. -Type the annotations and click ‘OK’. -A short description of the main observations is now shown on the the second graph. -Move the annotation to desired area so that it does not overlap with the bars.
-Under ‘Format’->‘Lines’, select None for row grid lines.
3.2.3. Create the third graph
1. create a new sheet.
-drag ‘Age Group’ to Columns and ‘LFPR(%)’ to Rows.
-drag ‘Years’ to Filters and select only year 2010 and 2021.
-change Marks to ‘Line’.
-drag ‘Years’ to ‘Path’.
-drag ‘LFPR(%)’ to Rows again and right click and check ‘Dual Axis’
Above changes will produce following graph:
-Set Marks as shape and drag ‘Years’ to Detail
The following graph is produced:
2. Create table calculation
-Create new calculated field called ‘Difference’
-Click ‘Analysis’->‘Create Calculated Field’ and fill in the formula shown in the picture below.
-Click ‘Default Table Calculation’ and set ‘Compute using’ as ‘Years’.
-Click ‘Ok’.
-Create new calculated field called ‘Direction’.
-The calculation depends on the value of ‘Difference’.
3. Update colors and shape
-Drag ‘Direction’ to ‘Color’ for both ‘SUM(LFPR(%))’
-Click on ‘Color’ and select ‘Edit Colors’ and change to the color to red and blue as shown below:
-Click on ‘Shape’ for second ‘SUM(LFPR(%))’ and select filled circle.
-Click on ‘Size’ to reduce the size of the circles to desired size.
4. Add in Labels
-Drag ‘Difference’ to ‘Label’ under the second ‘SUM(LFPR(%))’.
-Click on ‘Label’ and make the following changes:
-Change 'Text' to '<AGG(Difference)>%'.
-Under 'Font', choose 'Match Mark Color'.
-'Alignment' change to Top Right.
-Check 'Allow labels to overlap other marks'.-As there will be repeated labels on the graph, select all repeated labels by clicking on the labels while pressing the ‘ctrl’ key on the keyboard at at the same time. After selecting all the repeated labels,right click and select ‘Mark Label’->‘Never Show’.
Above changes will put labels on the graph as shown below:
5. Edit title, hide axis and adjust grid lines
-Update the title name as ‘% Change in Population from
2020-2021’. Align it to center, Format it as bold and set size to 11.
-Hide the right y-axis by right clicking on it and uncheck ‘Show Header’. Hide the x-axis in the same way.
-Under ‘Format’->‘Lines’, select None for row grid lines
-Select column grid lines and set the color to light grey.
-Hide ‘AGG(Direction)’ card on the right.
-Rename the sheet as ‘graph3’.
Following graph is produced:
3.2.4. Create dashboard
-Drag ‘graph1-2’ and ‘graph3’ onto the dashboard.
-Under ‘Dashboard’, set the ‘Size’ to 1000x600.
-Place ‘graph3’ at the bottom, adjust its size and align its vertical grid lines with ‘graph1-2’.
-Rename the dashboard as ‘LFPR’
The dashboard is now completed:
4. Main Observations
1. It can be observed that age groups having high LFPR are from 25-54. Each age group within this range has average LFPR above 80% over the year 2010-2021.
2. The age groups mentioned in point 1 also has low % change in yearly LFPR as shown by the low bar heights in the second graph. The overall change from 2010 to 2021 is also low, with highest percentage change of 6.20%.
Point 1 together with point 2 indicates that LFPR for working age within 25-54 is high over the years and the values are relatively stable too, suggesting stable increase in workforce population over the years.
3. Age groups for elderly, 65-69, 70-74 and 70 & over, are having increasing trend in LFPR over the years as observed through the increasing bar height for LFPR over 2010-2021. This corresponds to the second graph too, as most of the bars are of positive value, indicating an
increasing trend. Apart from positive bar values, the bar heights showing % changes are higher than age groups within range of 25-54, suggesting higher fluctuations. This corresponds to the percentage change values shown by the third graph, with 20% and 30.10% increase for age groups
65-69 and 70-74, and 6.40% increase for age group 70 & Over. This indicates that more and more elderly are participating in work, creating more labor force for Singapore as a result.
4. Looking at the third graph, LFPR for all age groups except 20-24 are increasing. This may be because of reasons such as more and more people pursuing higher education after graduation as indicated by the increasing number of graduates from higher degree courses link. However, it may of interest to user of the data to explore further the reasons behind it.
5. References
Link for data: (https://stats.mom.gov.sg/Pages/LabourForceTimeSeries.aspx)
Link for age group categorization : (https://www.singstat.gov.sg/-/media/files/standards_and_classifications/nsa.pdf)
Link for number of graduates from higher degree courses: (https://tablebuilder.singstat.gov.sg/table/TS/M850551)
The idea of third graph is from previous student’s work: (https://lizhenglong.netlify.app/posts/2021-01-27-dataviz-makeover-1/)